NanoGPT
Pay-per-prompt confidential inference — TEE-backed models with ECDSA per-request attestation
NanoGPT runs a curated set of open-source models inside Trusted Execution Environments (TEE), providing hardware-enforced confidentiality for prompts and responses. TEE models are identified by a TEE/ prefix and can be cryptographically verified: a dedicated attestation endpoint returns the hardware quote, and a per-request ECDSA signature endpoint lets clients confirm their specific completion was produced inside the enclave. No account or subscription required — anonymous cryptocurrency payments accepted.
API Features
OpenAI Compatible ✓
Streaming ✓
Function Calling —
Batch Inference —
Vision —
Embeddings —
Models & Pricing
| Model | Family | Context | Max Out | Input /M | Output /M | Free |
|---|---|---|---|---|---|---|
| Qwen3 235B A22B | Qwen | 250K | — | $0.13 | $0.50 | — |
| GPT-OSS 120B | OpenAI GPT | 128K | — | $2.00 | $2.00 | — |
| GPT-OSS 20B | OpenAI GPT | 128K | — | $0.20 | $0.80 | — |
| Qwen 2.5 VL 72B | Qwen | 64K | — | $0.70 | $0.70 | — |
| DeepSeek V3.1 | DeepSeek | 160K | — | $1.00 | $2.50 | — |
| DeepSeek V3.2 | DeepSeek | 160K | — | $0.50 | $1.00 | — |
| Qwen3 30B A3B | Qwen | 256K | — | $0.15 | $0.45 | — |
| GLM-4.6 | ZhipuAI GLM | 198K | — | $0.75 | $2.00 | — |
| GLM-4.7 Flash | ZhipuAI GLM | 198K | — | $0.15 | $0.50 | — |
| DeepSeek R1 0528 | DeepSeek | 125K | — | $2.00 | $2.00 | — |
| Gemma 3 27B | Google Gemma | 128K | — | $0.20 | $0.80 | — |
| Llama 3.3 70B | Meta Llama | 125K | — | $2.00 | $2.00 | — |
| Qwen3 Coder 480B | Qwen | 125K | — | $1.50 | $2.00 | — |
| Kimi K2 Thinking | Moonshot AI | 125K | — | $2.00 | $2.00 | — |
| MiniMax M2.1 | MiniMax | 195K | — | $0.30 | $1.20 | — |
| Kimi K2.5 | Moonshot AI | 125K | — | $0.30 | $1.90 | — |
| Kimi K2.5 Thinking | Moonshot AI | 125K | — | $0.30 | $1.90 | — |
| GLM-4.7 | ZhipuAI GLM | 128K | — | $0.85 | $3.30 | — |
| GLM-5 | ZhipuAI GLM | 198K | — | $1.20 | $3.50 | — |